Can pre-operative axial CT imaging predict syndesmosis instability in patients sustaining ankle fractures? Seven years' experience in a tertiary trauma center

The purpose of this study is to explore the diagnostic accuracy of CT measurements in predicting syndesmosis instability of injured ankle, with correlation to operative findings. From July 2006 to June 2013, 123 patients presented to a single tertiary hospital who received pre-operative CT for ankle fractures were retrospectively reviewed. All patients underwent open reduction and internal fixation for fractures and intra-operative syndesmosis integrity tests. The morphology of incisura fibularis was categorized as deep or shallow. The tibiofibular distance (TFD) between the medial border of the fibula and the nearest point of the lateral border of tibia were measured at anterior (aTFD), middle (mTFD), posterior (pTFD), and maximal (maxTFD) portions across the syndesmosis on axial CT images at 10 mm proximal to the tibial plafond. Statistical analysis was performed with independent samples t test and ROC curve analysis. Intraobserver reproducibility and inter-observers agreement were also evaluated. Of the 123 patients, 39 (31.7 %) were operatively diagnosed with syndesmosis instability. No significant difference of incisura fibularis morphology (deep or shallow) and TFDs was demonstrated respective to genders. The axial CT measurements were significantly higher in ankles diagnosed with syndesmosis instability than the group without (maxTFD means 7.2 +/- 2.96 mm vs. 4.6 +/- 1.4 mm, aTFD mean 4.9 +/- 3.7 mm vs. 1.8 +/- 1.4 mm, mTFD mean 5.3 +/- 2.4 mm vs. 3.2 +/- 1.6 mm, pTFD mean 5.3 +/- 1.8 mm vs. 4.1 +/- 1.3 mm, p < 0.05). Their respective cutoff values with best sensitivity and specificity were calculated; the aTFD (AUC 0.798) and maxTFD (AUC 0.794) achieved the highest diagnostic accuracy. The optimal cutoff levels were aTFD = 4 mm (sensitivity, 56.4 %; specificity, 91.7 %) and maxTFD = 5.65 mm (sensitivity, 74.4 %; specificity, 79.8 %). The inter-observer agreement was good for all aTFD, mTFD, pTFD, and maxTFD measurements (ICC 0.959, 0.799, 0.783, and 0.865). The ICC for intraobserver agreement was also very good, ranging from 0.826 to 0.923. Axial CT measurements of tibiofibular distance were useful predictors for syndesmosis instability in fractured ankles. The aTFD and maxTFD are the most powerful parameters to predict positive operative instability.